• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

Palit Jetstream 970 - first time overclocking

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

tatalor

New Member
Joined
Jan 8, 2016
Hello there,

First time trying to overclock anything here.

Using MSI Afterburner I managed to reach:

Power: 111
Core Clock +150 (1516Mhz)
Memory +500 (4000Mhz)

In fan profile I set the fan to 75% when temperature reaches 75C and 80% when temperature reaches 80C, and of course 100% when temp reaches 85C (I really hope the fan will stay at 75% most time).

Using Unigine Heaven I got:

FPS: 61.0
Score: 1536
Min FPS: 26.4
Max FPS: 128.7

Most of the time when running Heaven or playing, the temp hovers around 75-76C, but that's nothing new, it's about the same temp as non-overclocked using factory fan profile and in all reviews they said it was a hotter card than the rest.

Worked OK, about one hour running Heaven then half hour Rainbow Six and Crysis 3.

Then I had a blank screen after an hour or so of playing Call of Duty 3 (temp was about 78C) and after I turned off the computer from the power button (nothing else worked), Windows 10 Pro refused to load (just blank screening). Tried to system restore it and although at the end of it the message was System Restore failed, it actually worked.

Now all seems fine, had to reinstall Unigine Heaven, ran a couple of benchmarks on default, normal results. Wondering if was connected to the overclock or just an unlucky system crash. My hunch is that the overclock played some part :) Quite a scare! An hour before that I ordered Noctua NH-D14 thinking to overclock the CPU too (i5 Skylake 6600k) but now I'm unsure about it.

My values were quite high compared to the reviews I saw for this specific card, no idea if too high though.

What do you guys think? Would appreciate some opinions or advice.

Thank you in advance!
 
Last edited:
Welcome to the forums!

My guess is that your overclock most likely caused the crash.

You did not list your system specs...can your power supply handle the load from the card? Is it a quality power supply?

The 1516 MHz GPU clock...is that boost speed? Do you stay at that boost speed all the time during the Heaven benchmark run? The Heaven benchmark results don't really tell us much, unless you give your resolution and the quality settings you used when you ran it.

Depends on how close you want to live on the edge, but I would back off the GPU overclock to 100 MHz, and the memory to 250 MHz...let it run like that for a while (days/weeks). If it's stable, add 10 MHz to the GPU, 25 to the memory...let it run like that for a while (days/weeks). Repeat.
 
My PSU is a SeaSonic M12II 620 Bronze 620W, Gigabyte Z170X Gaming 3, i5 6600k Skylake (which I plan to overclock as soon as I get a new cooler) and 16Gb RAM (Corsair Vengeance LPX).

1516 Mhz is indeed boost speed and I think it stays at max during Heaven.

Heaven is maxed out including tessellation, resolution is 1920x1080.

I think downclocking a bit is a good advice and I'll follow it. Also, I upped the fan a bit and hopefully the temp will stay at 75C.
 
Depends on how close you want to live on the edge, but I would back off the GPU overclock to 100 MHz, and the memory to 250 MHz...let it run like that for a while (days/weeks). If it's stable, add 10 MHz to the GPU, 25 to the memory...let it run like that for a while (days/weeks). Repeat.
wow.. that would take weeks or months to find stability...yikes. :-/

Just loop heaven for a few minutes. When it locks or shows artifacts, back it down 10/20 MHz and go again. Once you find that middle ground, then loop heaven for a few hours for a final stability test.

There is, IMO, no need for such a prolonged process. :)
 
Then I had a blank screen after an hour or so of playing Call of Duty 3 (temp was about 78C) and after I turned off the computer from the power button (nothing else worked), Windows 10 Pro refused to load (just blank screening). Tried to system restore it and although at the end of it the message was System Restore failed, it actually worked.

^ This is enough reason :D
 
Sorry, a bad gpu overclock doesn't typically bork a windows install.
Ive never had that happen, even under ln2. I wouldn't bet my life the gpu overclock did it. Correlation is not causation!

Overkill.. days for stability testing??? To each their own but.... come on, lol!
 
+1 Video card overclock should not kill the OS data that would be a CPU overclock or bad hard drive.
 
I agree that a video card overclock "should" not kill an OS installation, but it can happen.

1. Video card driver crashes hard (i.e. system is locked up, no choice but to power cycle)
2. OS is writing something "important" when the system crashed hard
3. OS tries to boot, and it can't get important data
4. OS must be repaired or re-installed

I had this happen to me one time when I was first overclocking my first 970. Turned out that the card was bad (had to send it back on RMA), but the graphics driver crashed hard, crashed the OS. I had to power cycle. The OS was borked...had to reinstall from my last save backup.

Since then, I have been very cautious (as cautious as CPU overclocking) in how I approach my GPU overclock.
 
Yes I agree that is why I said it should not.:) It's just hard this day and age to see that MS would still have things writing to the kernel or boot sector wen running, like windows 95 you always needed to do a shutdown or it borked the OS sometimes..
 
Last edited:
I agree that a video card overclock "should" not kill an OS installation, but it can happen.

1. Video card driver crashes hard (i.e. system is locked up, no choice but to power cycle)
2. OS is writing something "important" when the system crashed hard
3. OS tries to boot, and it can't get important data
4. OS must be repaired or re-installed

I had this happen to me one time when I was first overclocking my first 970. Turned out that the card was bad (had to send it back on RMA), but the graphics driver crashed hard, crashed the OS. I had to power cycle. The OS was borked...had to reinstall from my last save backup.

Since then, I have been very cautious (as cautious as CPU overclocking) in how I approach my GPU overclock.
That's too cautious... by leaps bounds, overkill nuclear weapon in place of a firecracker.


Days of stability testing.... chuckles... :)
 
That's too cautious... by leaps bounds, overkill nuclear weapon in place of a firecracker.


Days of stability testing.... chuckles... :)

To each is own! :D

I was new to overclocking when the GPU crash borked my Windows install...freaked me out.

I run my machine 24/7 doing...something! :D So super stability is my main driver!
 
In the end I went with +130 core clock (1496Mhz boost) and +350 memory (3850Mhz). All in all I think it's a pretty good overclock and seems stable after running Heaven about 10 times and playing couple hours Call of Duty 3. Also I upped to fan to 85% which is a bit noisy for my taste, but at least the temperature very rarely goes over 75C. Maybe I'll up it a bit more if all is fine after some time. And, by the way, which is more important to overclock, core clock or memory?

Another question. Since I use a 32" Sony TV as a monitor and I'm limited to 60Hz / 60 FPS no matter what, is there a problem if I switch overclocking on and off depending if I need it? When I'm already at 60FPS with top settings, there's no need for it and I'd rather have the fan spinning slower and be quiet. Would the card by affected by constantly turning overclocking on and off?
 
I agree that a video card overclock "should" not kill an OS installation, but it can happen.

1. Video card driver crashes hard (i.e. system is locked up, no choice but to power cycle)
2. OS is writing something "important" when the system crashed hard
3. OS tries to boot, and it can't get important data
4. OS must be repaired or re-installed

I had this happen to me one time when I was first overclocking my first 970. Turned out that the card was bad (had to send it back on RMA), but the graphics driver crashed hard, crashed the OS. I had to power cycle. The OS was borked...had to reinstall from my last save backup.

Since then, I have been very cautious (as cautious as CPU overclocking) in how I approach my GPU overclock.

I do know that it's possible to have a domino-effect with a sagging +12V rail. (if the PSU is having a problem)
 
And, by the way, which is more important to overclock, core clock or memory?

Core overclock will give you way more benefit than the memory overclock.

Another question. Since I use a 32" Sony TV as a monitor and I'm limited to 60Hz / 60 FPS no matter what, is there a problem if I switch overclocking on and off depending if I need it? When I'm already at 60FPS with top settings, there's no need for it and I'd rather have the fan spinning slower and be quiet. Would the card by affected by constantly turning overclocking on and off?

No problem with that at all. I don't remember you saying what OC software you used (I use the EVGA Precision X16). You can use different presets there and change the OC. The only thing I have noticed is that if I change the OC when the card is running folding@home, I will sometimes get a graphics driver crash.

I do understand the fan noise...I remember going through that when I was overclocking the 970s I used to have...it's one of the reasons I went to the 980 Ti hybrid (as I'm too much of a wimp for a custom water loop - hehe).
 
Core overclock will give you way more benefit than the memory overclock.

Only up to a certain point. After about 1600 to 1650 on the core for Maxwell, you'll get really diminishing returns. That with the GM204. With the GM200, that's likely closer to 1550ish. The memory, however, will continue to give good performance bumps all the way up to 2150 and above. I've only been able to run mine up to 2175, but.....the performance gains are there.
 
Back